Honey Bee Conservation Using Deep Learning

Explore the development of DeepBee, an innovative tool for automatic honeybee colony assessment, in this 44-minute conference talk from MLCon. Learn how deep learning techniques are revolutionizing bee conservation efforts by automating the traditionally laborious and error-prone task of counting and classifying comb cells. Discover the image classification pipeline that combines classical image processing methods with state-of-the-art Deep Neural Networks for image segmentation and classification. Follow the speakers' journey from image collection to final solution delivery, gaining insights into the comparison of 13 distinct DNN architectures and the selection process for the best model. Understand the challenges faced, mistakes made, and valuable lessons learned during the project's development at the Polytechnic Institute of Bragança. Gain a comprehensive understanding of how cutting-edge technology is being applied to support honey bee conservation and improve beekeeping practices.